ZReverend, bayesian classifier for Zope 3

Reverend is a lightweight python tool by Amir Bakhtiar that provides a bayesian classifier. It's a perfect tool to classify spam mails for instance.

http://www.divmod.org/projects/reverend

It was its first use case anyway, since the authors based it on Christophe Delord PopF tool:

http://christophe.delord.free.fr/

I need this tool for the webmail I am coding but since I have thaught of many other usages in zope for a bayesian classifier, I have created a small Zope 3 utility (very early stage right now).

The Z3 product is called "zreverend" and wraps reverend into zope by:

• automatically saving the bayesian network into persistent dictionnaries in the ZODB
• publish all useful methods to use reverend in some code
• provide a minima managment screen
• ... more to come i guess
  

That's my first "pure" zope 3 product so besides its usage it provided me a good training. (it still needs some good code review though)

Notice that Zope 2 developers can use it through Five, by manually instanciate ZBayes objects, like I did in CPSMailAccess with zemantic and like I will with zreverend.

Browse the code here:

https://svn.nuxeo.org/trac/pub/browser/z3lab/zreverend/trunk/

get the code with svn like this:

svn co https://svn.nuxeo.org/pub/z3lab/zreverend/trunk/ zreverend


If this tool is somehow useful to you or if please let me know !

(Post originally written by Tarek Ziadé on the old Nuxeo blogs.)